Content
82%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A tight, highly actionable skill body with executable examples and well-labeled one-level-deep references that exist in the bundle. Minor issues — a duplicated invocation line, a missing post-creation verification step, and broken eval scenario links — keep it just below top marks.
Suggestions
Remove the duplicated invocation instruction in Quick Start — merge "Invoke it in OpenCode: /test" and the "Run bun run opencode to start..." line into a single step, ordered start-then-invoke.
Add an explicit validation step after command creation, e.g., "Verify: run /help (or type /<name>) to confirm the command is registered and test it on a trivial input before iterating."
Fix or remove the Eval Scenarios links — evals/scenario-0..2/task.md do not exist in the bundle, so either include those files or drop the links from the distributed SKILL.md.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes competence ("NEVER use $ARGUMENTS more than once... WHY: The placeholder is substituted literally"), but the Quick Start repeats the invocation instruction — "Invoke it in OpenCode: /test" followed immediately by "Run bun run opencode to start, then type /<name> to invoke your command" — a minor instance of over-explanation that could be trimmed. This fits the score-4 anchor better than score 5, where every token would earn its place. | 4 / 5 |
Actionability | Fully executable, copy-paste-ready guidance: a complete command file with frontmatter and body ("Run the full test suite, identify any failures, and fix them. 1. Run `npm test`..."), concrete placeholder syntax ($ARGUMENTS, $1/$2, !cmd, @filename), and BAD/GOOD example pairs for every anti-pattern. Common cases are covered, matching the score-5 anchor. | 5 / 5 |
Workflow Clarity | The Quick Start sequence (create `.opencode/commands/<name>.md` → invoke `/<name>`) is clear and the command/agent/skill decision rule is explicit, but there is no verification step (e.g., confirm the command appears in /help or test it after creation), and the invocation instruction appears before the "run bun run opencode to start" step. This matches the score-4 anchor (clear sequence, minor validation gaps) rather than the explicit-validation score-5 anchor. | 4 / 5 |
Progressive Disclosure | Good structure: the References section clearly signals two one-level-deep files with content descriptions ("Creation phases, example commands... " and "Complete frontmatter fields and placeholder reference"), both of which exist in the bundle. However, the "Eval Scenarios" links point to evals/scenario-0..2/task.md, which do not exist in the bundle — a minor organization gap that keeps this below the score-5 anchor's "easy navigation". | 4 / 5 |
Total | 17 / 20 Passed |